Search within Lanny's blog:

Leave me comments so I know people are actually reading my blogs! Thanks!

Monday, January 19, 2009

My Research: BYU UAV Demo Dry Run

Hi, everyone who reads my blog! Happy New Year to all of you. Wish you a very exciting and productive new year! (See picture of the day below!) I only have one New Year resolution this year -- that is, to catch up with the blog! :) That means I'll have to post at least two blog entries each day! So get ready for a flood of interesting (and hopefully insightful) postings. Also be prepared for the strange parallel time/space I'll be living in.

Note that I am starting a new track today called "My Research." Postings with this tag will talk about AI/Robotics research I am working on. Hope you find inspirations in these postings and comments are especially welcome for this track!

If you have not noticed, there's a section on the right side of my blog called "Blog Labels." This is a good way to filter out postings you might find interesting to you. For example, there's a label for each book I translate. The only drawback is that you'll have to read backwards. :) Also if you like my postings, please follow my blog (see right side). I am interested to see how many people really like my postings, and the more people liking my postings, the more motivated I will be! Okay, enough babbling, let's move on to the real fun stuff.


Part of my research is about how to use an Unmanned Aerial Vehicle (UAV) to support Wilderness Search and Rescue (which we refer to as WiSAR). On November 14, 2009, our research group performed a field dry run in Elberta, Utah (a place in the middle of nowhere) in preparation for an upcoming demo for the Utah County Search and Rescue people.

Utahans love outdoor activities because we are blessed with lots of beautiful mountains and wilderness. As a side effect, there's also a great demand for wilderness search and rescue because people get injured/lost/missing in the wilderness. The goal of our research group is to use UAV technologies to support wilderness search and rescue operations. Obviously, real-time video from a UAV with bird's eye view can provide useful information for the search and rescuers, especially for areas that are hard to reach quickly. The UAV can also cover an area much faster than search and rescuers on foot. Our research group has been working on this for several years and made good progress. However, the technologies will only be able to make a difference if the search and rescuers find them useful and start using them. That was the reason why we are eager to do a demo for the real search and rescuers. And the purpose of the dry run is to make sure all technology components are ready.

The previous day's weather forecast predicted snow in the next day. Sure enough, when I left home at 7:30am, the ground was covered by snow. Elberta is about one hour drive from BYU campus. Interestingly, weather got better and better as I drove, and by the time I arrived at Elberta at 9:00am, there was no snow!

Elberta, Utah, early morning!

For our research, we use a fix-wing propeller-powered model-plane kind of UAV shown in the picture below. We also have a nice trailer, which has a power generator, some mounted LCD monitors, a long table, and even a microwave!

Fix-wing UAV and its container

Outside look of the trailer (showing the power generator)

Inside view of the trailer

It took about 30 minutes to get everything set up. Meanwhile, an umbrella (marking the location of the missing person) had also been placed in a distance from the "command post." By 9:45am, we were ready to throw the plane into the air (literally, that's how we launch the UAV, because the UAV has built-in intelligence for auto-launching).

Ready? Launch!

Inside the trailer, we have two laptops running. One laptop is used to control the UAV with a program called Phairwell (don't ask me. I didn't pick the name), where the operator can set waypoints for the UAV to follow (or a flight pattern). The operator can also control the UAV's heading, speed, roll/pitch/yaw, height above ground, altitude, etc., etc. Another laptop is used to view video feed coming down from the UAV. It is worth mentioning that the video frames are actually mosaiced together so the video observer can view a larger area while each video frame stays on the screen for an extended time for the ease of searching.

Laptop running the UAV control software Phairwell

Laptop running video mosaicing software

Amazingly, the weather turned into something perfect! There's nothing more we could have asked!

Sunny Elberta! What a beautiful day!

The dry run was quite successful. We performed several flights and fixed a few glitches, especially with the auto landing control. The picture below shows how the UAV lands (yes, it's a crash landing). The picture was actually taken from a previous field trial because it is quite difficult to try to keep the UAV in the camera frame.

UAV auto-landing

At 11:30am, just when we were ready to enjoy our lunch (subway sandwiches) after a successful dry run, guess what, it started to snow!!


We ended up packing everything first, and then had our lunch inside the trailer (aren't we glad there is a microwave in the trailer!). There's me packing in the snow in the picture below. Don't ask me why those other two were doing synchronized penguin walk in the background, cause I don't know!

That's it! We were fortunate enough to have a window of nice weather (against the weather forecast prediction) for the dry run, and we were ready for the demo!!

See the complete gallery for the dry run
Download geo-tagged photos for Google Earth view (double click the kml file)

Picture of the Day:

Wish you all a very exciting New Year! Hee-Ha!